Adding zoom on scrollwheel event to Pool demo.

Adding handlers to zoom in/out on scrollwheel event.

Questions:
- Is deltaY the correct field in which to store this event value?
- Is the 'wheel' addEventListener hook broken in the first place? Doesn't work for me on OSX, but might be working in Windows/Linux and don't know if this change would break it for those.
- It is necessary to cancel event bubbling for the onmousewheel event?

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=49305

[petersont, 2010-06-02 01:27:46]

Brilliant change.  deltaY is correct, and I think I recall that the different
behavior of mouse events in the plugin was the reason I didn't do this in the
pool game in the first place.  So, LGTM aside from one comment.

http://codereview.chromium.org/2456004/diff/4001/5002
File samples/o3d-webgl/client.js (right):

http://codereview.chromium.org/2456004/diff/4001/5002#newcode854
samples/o3d-webgl/client.js:854: event.deltaY = -1 * info.wheel;
Why not just -info.wheel?

[luchen, 2010-06-02 01:49:13]

+kbr to reviewer list; seeking feedback on event propagation: should mousewheel
events in webgl context be passed up the chain, or should we cancel bubbling?

http://codereview.chromium.org/2456004/diff/4001/5002
File samples/o3d-webgl/client.js (right):

http://codereview.chromium.org/2456004/diff/4001/5002#newcode854
samples/o3d-webgl/client.js:854: event.deltaY = -1 * info.wheel;
On 2010/06/02 01:27:46, petersont wrote:
> Why not just -info.wheel?

Done.

[Ken Russell (switch to Gerrit), 2010-06-02 22:32:20]

On 2010/06/02 01:49:13, luchen wrote:
> +kbr to reviewer list; seeking feedback on event propagation: should
mousewheel
> events in webgl context be passed up the chain, or should we cancel bubbling?

If the behavior without canceling bubbling is that the page will scroll in
addition to the scroll event being dispatched to the O3D application, then I
think we should cancel bubbling.

[luchen, 2010-06-04 22:00:09]

That's the current behavior: your custom callback will fire, in addition to the
page scrolling (if there's a scrollbar).

The other event types execute this way though (e.g. rotatemodel.html has a
keyevent handler and explicitly adds a o3djs.event.cancel(ev) in the
user-written custom callback).
So perhaps it is worth *not* canceling bubbling, in order to be consistent with
the other event types. Thoughts?


On 2010/06/02 22:32:20, kbr wrote:
> On 2010/06/02 01:49:13, luchen wrote:
> > +kbr to reviewer list; seeking feedback on event propagation: should
> mousewheel
> > events in webgl context be passed up the chain, or should we cancel
bubbling?
> 
> If the behavior without canceling bubbling is that the page will scroll in
> addition to the scroll event being dispatched to the O3D application, then I
> think we should cancel bubbling.

[Ken Russell (switch to Gerrit), 2010-06-04 22:08:53]

I can't imagine a situation where if you had a scroll event handler installed
that you'd also want the page to scroll. I leave it to your discretion, though
it sounds like we will either want the canceling to happen automatically, or in
each demo that uses a scroll wheel event.

On 2010/06/04 22:00:09, luchen wrote:
> That's the current behavior: your custom callback will fire, in addition to
the
> page scrolling (if there's a scrollbar).
> 
> The other event types execute this way though (e.g. rotatemodel.html has a
> keyevent handler and explicitly adds a o3djs.event.cancel(ev) in the
> user-written custom callback).
> So perhaps it is worth *not* canceling bubbling, in order to be consistent
with
> the other event types. Thoughts?
> 
> 
> On 2010/06/02 22:32:20, kbr wrote:
> > On 2010/06/02 01:49:13, luchen wrote:
> > > +kbr to reviewer list; seeking feedback on event propagation: should
> > mousewheel
> > > events in webgl context be passed up the chain, or should we cancel
> bubbling?
> > 
> > If the behavior without canceling bubbling is that the page will scroll in
> > addition to the scroll event being dispatched to the O3D application, then I
> > think we should cancel bubbling.

[luchen, 2010-06-08 18:31:38]

Good point. The O3D plugin implementation cancels the scroll event automatically
for you, so I did the same here. PTAL.

 On 2010/06/04 22:08:53, kbr wrote:
> I can't imagine a situation where if you had a scroll event handler installed
> that you'd also want the page to scroll. I leave it to your discretion, though
> it sounds like we will either want the canceling to happen automatically, or
in
> each demo that uses a scroll wheel event.

[Ken Russell (switch to Gerrit), 2010-06-08 18:34:30]

LGTM

[petersont, 2010-06-09 19:01:03]

LGTM.